I had a minor crash today and I would greatly apreciate any help help I can get from you guys:)!
I have been flying this setup without problems with about 60 minutes of total air time experience. I have been flying in stabilise, alt hold and loiter without any problems. Loiter and alt hold are very stabile.
Today I took off in stabilise, then switched to alt hold and then loiter. Everything was going fine. I flew around in loiter mode, where the sticks simply control direction while maintaining altitude. When I was about to land I was still in loiter and brought the quadcopter down to about 4 meters, still in loiter. Then I thought I switched to alt hold but from the log I can see that I switched to stabilise which should also be fine. However the moment I switched to stabilise the all motors stopped and the copter dropped from the sky. My throttle was a bit above middle which should be fine to hover. My copter hovers fine at around 50% throttle. From the logs I can see that throttle out is reduced to almost zero when it enters stabilise mode. Throttle in is about 50% you can see that I raise throttle as the copter drops. When it hit the ground I lowered throttle to zero.
Any help is greatly appreciated:)! Below you can see graph of throttle in, throttle out, voltage and current. I do not have any low voltage alarms or cut-off enabled. Logs are also attached. I removed the beginning of the flight in the .log file since it was larger than the 7mb limit.
I'm using Arducopter 3.0.1 release.
The problem isn't that the motors stopped it is that you had an un-commanded roll to the left. As you go past 90 degrees the motors go to minimum to stop you from accelerating hard into the ground.
Is this a hex?
You have removed the beginning of the log and that takes a lot of information from us that we use to understand the problem.
Yes, I agree with Leonard. It's most likely a mechanical failure which causes it to lose roll control. To see it you need to graph the NTUN message's DRoll and DPitch (desired roll and desired pitch) which are the right-most columns in the NTUN message. The graph the ATT message's Roll and Pitch (i.e. copter's actual roll and pitch).
They desired and actual stick together pretty well right up until the last moment and then suddenly diverge. It might seem a strange coincidence that it happens as you switch to stabilize mode but actually it goes horribly wrong before the switch to stabilize (about 8/10ths of a second before). So the copter is already leaning over at 95 degrees to the left when stabilize mode is engaged.
Impressive flight time though! I looked at the log files and thought, "jeepers, are multiple flights being concatenated? These logs are huge!".
Thank you very much for your input leonardthall:)!
It did not occur to me that this was a mechanical problem. I can see from my complete log that it does not have the settings in the beginning either. Maybe it got too big and it started deleting the beginning?(10.1MB) However I have attached another flight log with the settings in it. Hopefully this is useful. All settings should be the same in this log.
It's a quad with 17" props and AUW of 2.200grams
Any further findings would be awesome:)! I will look into the mechanical parts today and get back to you. Yesterday was concentrated on looking at the logs.
Yes, when the logs fill up they start back from the beginning but the dataflash logger sticks the FMT (format) messages in there every now and then just in case this kind of things happens.
Cool! That makes sense. Thanks for explaining:)!
Thanks for your reply. It really helps a lot that you took your time to explain how and where in the logs I can see the problems. Hopefully in time I won't need as much help the experts and maybe even be able to help other people:)!
As I mentioned to leonardthall I will look through the mechanics today to see if I can find a failure.
I also have onboard 720p at 100fps video from the crash. I will slow it down and try to embed here. Maybe it will shed more light on what happened.
I will post back when I have more information. Again thank you very much for your excellent help!
Here is the youtube video of the crash in normal followed by slow motion. I can't hear if it is one motor cutting or many. Maybe you guys have better ears:)? It does seem like it falls quite evenly to the left though.
I just wanted to let you know that I believe I have found the issue. One of the wires going from ESC to motor was broken. So just as you mentioned it was indeed a mechanical failure. There was a bit strain on this wire which probably caused it to break during vibration and wear. I will recheck all wires and solder extensions if needed. The esc was the back left one(2 in X setup) which corresponds fine with the video where the quad drops to the left side. Here is a picture of the wire break. Thanks again leonardthall and Randy! Your help is greatly appreciated:)!